Method forward

semantic_sam/backbone/focal_dw.py:168–204

Forward function. Args: x: Input feature, tensor size (B, H*W, C). H, W: Spatial resolution of the input feature.

(self, x)

            self.gamma_2 = nn.Parameter(layerscale_value * torch.ones((dim)), requires_grad=True)

    def forward(self, x):
        """ Forward function.

        Args:
            x: Input feature, tensor size (B, H*W, C).
            H, W: Spatial resolution of the input feature.
        """
        B, L, C = x.shape
        H, W = self.H, self.W
        assert L == H * W, "input feature has wrong size"

        x = x.view(B, H, W, C).permute(0, 3, 1, 2).contiguous()
        x = x + self.dw1(x)
        x = x.permute(0, 2, 3, 1).contiguous().view(B, L, C)

        shortcut = x
        if not self.use_postln:
            x = self.norm1(x)
        x = x.view(B, H, W, C)
        
        # FM
        x = self.modulation(x).view(B, H * W, C)
        x = shortcut + self.drop_path(self.gamma_1 * x)
        if self.use_postln:
            x = self.norm1(x)

        x = x.view(B, H, W, C).permute(0, 3, 1, 2).contiguous()
        x = x + self.dw2(x)
        x = x.permute(0, 2, 3, 1).contiguous().view(B, L, C)

        if not self.use_postln:
            x = x + self.drop_path(self.gamma_2 * self.mlp(self.norm2(x)))        
        else:
            x = x + self.drop_path(self.gamma_2 * self.mlp(x))
            x = self.norm2(x)

        return x

class BasicLayer(nn.Module):
    """ A basic focal modulation layer for one stage.